I almost missed it. It came out on June 16th, buried under a wall of Google spam update coverage, and I only caught it because someone posted a screenshot in a Slack I'm in. The caption said something like "wait, Bing actually did something useful?" and yeah — they really did.
Microsoft quietly added four new capabilities to Bing Webmaster Tools under the AI Performance section: Intents, Topics, Citation Share, and Compare. Free. In preview globally. No enterprise contract, no third-party data broker, no scraping prompts at 2am. Just go to Bing Webmaster Tools and it's there waiting for you.
The one that matters most is Citation Share. And I want to explain exactly why, because if you're doing any kind of generative engine optimization work in 2026, this is the measurement gap that has been making everyone's life miserable.
The Measurement Problem That's Been Killing GEO Work
Here's the frustrating reality of working in AI search visibility: until now, you were mostly flying blind. You could check if your site showed up in a Perplexity answer. You could run prompts manually and screenshot results. Some agencies were literally paying people to submit hundreds of prompts and tally citations in spreadsheets.
The problem is that AI systems don't show you a "you were cited 847 times this week" dashboard. They just answer questions. Whether your content is in the mix or not is usually invisible unless someone happens to ask the right thing and your page floats to the top.
This is what made the announcement from Krishna Madhavan at Microsoft interesting. Citation Share isn't just "how many times were you cited" — it's "what percentage of the citation space did you capture for a specific grounding query." That's a different, more useful signal. If you're getting cited 40 times but there are 200 total citations across all sites for that query, you have a 20% Citation Share. That tells you something actionable: you're in the conversation, but you're not dominating it.
What the Four New Features Actually Do
Let me go through each one because they work together.
Intents
This classifies your grounding queries by intent type — informational, comparison, transactional, etc. Why does this matter? Because if 80% of your AI citations are coming from comparison-intent queries but you thought you were winning on informational queries, you need to know that. Your content strategy and your actual AI visibility are probably misaligned.
Topics
Instead of looking at individual queries one at a time, Topics groups your citation activity into broader thematic areas — the same structure that AI systems use to organize information internally. This is significant because modern language models don't think in keywords; they think in topics and entities. If Microsoft is showing you data organized that way, you're getting insight closer to how the system actually evaluates your content.
Citation Share
The headliner. For any grounding query where your content appears, Citation Share tells you what fraction of the total citation pool your site captured. It's designed to be an observational metric, not a ranking score — Microsoft is careful to say it's not a competitive scoreboard. But directionally, watching Citation Share trend upward after you publish new content tells you something real about whether your GEO work is working.
Compare
Date range comparison for citation activity. Simple but necessary. You can now look at how citations changed after a piece of content went live, or after a major AI update, or after a competitor entered the space. Before this, you were guessing.
Why This Beats What Google Currently Offers
This is the part that's genuinely surprising. Google has the AI Performance Report in Search Console — rolled out to the UK first, then progressively. It shows you impressions from AI Overviews and AI Mode. But impressions is a blunt instrument. It tells you how often your content was surfaced, not how you stack up against other sources surfaced for the same query.
Bing's Citation Share is more nuanced. It gives you a relative view. And as of right now, Google has nothing equivalent. There's no "here's your share of the AI citation space for [query]" report anywhere in Search Console. Not yet.
Now, I know what you're thinking: who cares about Bing? Fair question. Bing's market share in traditional search is what it's always been — the browser you use to download Chrome. But Microsoft Copilot is embedded in Windows, in Edge, in Teams, in Word, in Excel. The number of people getting AI-generated answers in a Microsoft product without ever opening a browser is not nothing. And the infrastructure Microsoft is building here — the citation tracking, the intent classification — is a preview of what the industry's measurement standards are going to look like when Google eventually builds the same thing.
In other words: starting your GEO measurement practice in Bing right now means your team knows how to interpret citation data by the time Google ships equivalent tools. That's worth something.
🤖 Check Your AI Search Visibility Right Now
See how your site performs across AI search engines — not just traditional rankings. RankSorcery's AI Search Visibility tool checks whether your content is positioned to get cited in ChatGPT, Perplexity, Bing Copilot, and Gemini.
Check My AI Visibility →How to Actually Use Citation Share (Step by Step)
Alright, enough context. Here's what to actually do with this data if you want to move the needle.
Set Your Baseline Now
Go to Bing Webmaster Tools and pull the AI Performance report for the last 30 days. Screenshot it. Export it. The Compare feature is only useful if you have historical data to compare against, and that data only starts accumulating from when you first check. Don't wait another week — you're burning baseline data every day you delay.
Find Your Strongest Grounding Queries
Sort by total citations first. These are the queries where AI systems are already pulling from your content. Now look at Citation Share for those same queries. If you have high citation volume but low Citation Share, there's a content coverage issue — other sources are getting pulled alongside you more often. That's your first optimization target.
Review Your Topics Distribution
Look at which Topics are driving your citation activity. Compare that to what you think your site is authoritative about. If they don't match, you have a topic coverage gap or a content structure problem. AI systems are telling you how they categorize your expertise, and that's not always what you intended.
Check Intent Mix Against Your Goals
If you're an e-commerce site and most of your citations are informational rather than comparison or shopping-intent queries, that's a signal your product pages aren't getting surfaced in the buying-decision moments. That has direct revenue implications and tells you what types of content to build next.
Track Before and After Content Updates
Pick two or three pages you want to improve for AI visibility. Update them with clearer entity definitions, better structured data, more direct factual statements, and improved coverage of related concepts. Then use the Compare feature to see if Citation Share moves for the grounding queries those pages are associated with. This is how you build an actual feedback loop instead of just guessing.
What Content That Gets High Citation Share Actually Looks Like
I want to get into the specific structural things that seem to drive higher AI citation rates, because "make good content" is not useful advice.
- Direct declarative statements near the top. AI models tend to pull content that has a clear, quotable answer early in the page — not buried in paragraph six after three hedging sentences.
- Specific numbers over vague claims. "Conversion rates dropped 18% in Q2" gets cited over "conversion rates fell significantly." Specificity signals primary research.
- Clear entity relationships. Define your subject, its relationship to adjacent concepts, and who it affects. Structured definitions get pulled more reliably than narrative prose.
- Updated timestamps with substantive changes. AI systems pay attention to freshness signals. Updating a last-modified date without changing content is a bad move — they can tell.
- FAQ sections with tight Q&A pairs. Not the old-style SEO FAQ with keyword-stuffed questions, but actual natural-language questions answered in one or two sentences. These are grounding-friendly by structure.
- Authoritative attribution. Content that cites its own sources — studies, official data, named experts — appears more credible to grounding systems than unsourced assertions.
- Minimal navigation clutter around the main content. AI scrapers have gotten better, but pages where the main content is clearly delineated from chrome still process more cleanly.
The Limits You Should Know Before Over-Interpreting This Data
Microsoft is clear that this is in preview, and there are real gaps. Citation Share only covers Bing and Copilot — it tells you nothing about ChatGPT, Perplexity, Google AI Mode, or Claude. So if Bing isn't a significant source of AI traffic for your audience, the absolute numbers will be small.
More importantly: Citation Share is observational. It doesn't tell you whether being cited translated to a user clicking through to your site. AI answers frequently use content as grounding without ever surfacing the URL to the user. High Citation Share might mean high influence over AI answers in your niche, or it might mean your content is being consumed without generating any visits. You need to correlate with your own analytics to tell the difference.
The Compare feature also only works if you have enough historical data. If you've never looked at your AI Performance report before, you have no baseline. Whatever Citation Share you set up this week is day one. So start now — but understand that your first month of data is context-building, not analysis.
Where This Is All Heading
I think what Bing did here is signal something important about where GEO measurement is going. The era of "we'll figure out AI visibility when Google gets around to giving us data" is ending. Publishers who wait for Google to build the full citation-level reporting they need are going to be months behind the people who built their measurement practice in Bing first.
The deeper point is this: AI search visibility is increasingly its own discipline, separate from traditional rankings. Your position-3 ranking tells you almost nothing about how often your content gets pulled as a grounding source. Those are different systems, different signals, different optimization patterns. And the measurement tools for AI visibility are just beginning to become available.
Bing jumped first. Pay attention to what they built, because Google's version is coming — and when it does, you'll either be ready to use it, or you'll be reading someone else's recap about why you should have started earlier.
Open Bing Webmaster Tools. Find AI Performance. Pull the last 30 days. That's all you need to do today.